Alternative treatment approaches are urgently needed to improve quantity and quality of life of patients with metastatic castration resistant prostate cancer (mCRPC). Immunotherapy in the form of checkpoint inhibition has yielded auspicious results in many cancers, including lung cancer, melanoma, renal cell carcinoma, and lymphoma amongst several others. The frequency of tumor-specific somatic mutations and hence neoantigen formation strongly predicts for objective response to checkpoint inhibitors (CPIs). Prostate cancer is typified by a relatively low mutational burden, so it is not surprising CPIs for mCRPC are largely ineffective, although occasional responses have been observed. Over the last few years, mismatch repair deficiency (dMMR) and biallelic inactivation of CDK12 (CDK12-/-) have been observed in a small subset of mCRPC patients. Importantly, these genetic lesions have been associated with increased mutational burden due to increased point mutations in the case of dMMR and heightened formation of focal tandem duplications in the case of CDK12-/-. Accordingly, dMMR or CDK12-/- tumors are expected to be sensitive to CPIs. The frequency of dMMR has been underestimated due to poor sensitivity of detection assays, which can fail to detect allelic inactivation of MMR genes. Moreover, neither the frequency of dMMR or CDK12-/- nor the response to checkpoint inhibition or dMMR or CDK12-/- tumors has been studied amongst Veterans. The relevance of this knowledge gap amongst Veterans is highlighted by key demographic differences in Veterans compared to the US population at large, which could in principle affect the frequencies of dMMR and CDK12-/- as well as response to checkpoint inhibition. Using a sensitive next generation sequencing platform and bioanalytic tool to detect microsatellite instability, a surrogate for dMMR, we predict that we can detect dMMR or CDK12-/- in at least 15% of Veterans. Furthermore, we hypothesize that Veterans with dMMR or CDK12-/- will exhibit a high response rate to checkpoint inhibition. We propose three aims: 1. Identify the frequency of dMMR and CDK12-/- as determined by NGS and a sensitive analytic tool for MSI detection amongst Veterans with mCRPC. The MSI bioanalysis, known as mSINGS (microsatellite instability by next generation sequencing), as well as targeted gene sequencing inclusive of CDK12 are built-in components of OncoPlex, one of the CLIA-certified NGS platforms being implemented within the VA Precision Oncology Program Cancer of the Prostate (POPCAP) network. 2. Perform an open label phase 2 clinical trial of pembrolizumab, an anti-PD1 CPI, to determine the efficacy of pembrolizumab amongst dMMR and CDK12-/- mCRPC Veterans who have received prior AR signaling ? inhibitors. The primary endpoint will be response rate, defined as a composite of: objective response rate by iRECIST 1.1, PSA50 at 12 weeks, radiographic progression free survival at 6 months. 3. Perform exploratory analyses to identify biomarkers of response and resistance to pembrolizumab in dMMR and CDK12-/- mCRPC Veterans. For this purpose, we will acquire baseline and at-progression biopsies of metastases as well as serial blood samples to investigate genomic, transcriptomic, proteomic, and immune factors that contribute to response. This study will result in enhanced understanding of the biology of mCRPCs amongst Veterans and the identification of Veterans who respond to checkpoint inhibition, thereby benefiting from potentially life extending therapy. In addition, exploratory/correlative analyses could identify molecular targets, the modulation of which could potentiate the anti-tumor effects of checkpoint inhibition..